common/strtol.cc: Get error testing also to work on FreeBSD

[wjwithagen]

• And report the same error types.
• Changed to report for the last error since the value is there but
  not allowed characters follow.

Error found by: run-cli-tests, because the wrong string was returned.

Signed-off-by: Willem Jan Withagen wjw@digiware.nl

[tchaikov]

can we just put

if (endptr == str && ret == 0) {
 ...

see http://pubs.opengroup.org/onlinepubs/9699919799/functions/strtol.html,

If the subject sequence is empty or does not have the expected form, no conversion is performed; the value of nptr shall be stored in the object pointed to by endptr, provided that endptr is not a null pointer.

and

Upon successful completion, these functions shall return the converted value, if any. If no conversion could be performed, 0 shall be returned [CX] [Option Start] and errno may be set to [EINVAL]. [Option End]

[wjwithagen]

@tchaikov
Could be that the FreeSD version is not 100% compliant...
Reason for this change is that in a test the "wrong" error messages was returned.
An that was because ret == 0 and errno == EINVAL Which triggered the first case.
Linux however chooses to not set errno != 0 but has str == endptr and selects the second case.
FreeBSD does indeed also set:

If there were no digits at all, however, strtoll()
     stores the original value of nptr in *endptr.

So perhaps just reversing the tests would already work.
I'll test that and adjust the PR if that works.

[wjwithagen]

@tchaikov

The following tests passed:
        run-cli-tests

Will amend PR

[tchaikov]

/home/jenkins-build/build/workspace/ceph-pull-requests/src/osd/ECBackend.cc: 2134: FAILED assert(res.returned.size() == to_read.size())

 ceph version 11.0.2-1728-gd9753d7 (d9753d7c653a9e49f70bf33518579ee4173f2909)
 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0x8b) [0x7f6654c70f2b]
 2: (CallClientContexts::finish(std::pair<RecoveryMessages*, ECBackend::read_result_t&>&)+0x566) [0x7f66548e7fa6]
 3: (ECBackend::complete_read_op(ECBackend::ReadOp&, RecoveryMessages*)+0x6f) [0x7f66548c7b4f]
 4: (ECBackend::handle_sub_read_reply(pg_shard_t, ECSubReadReply&, RecoveryMessages*)+0x1010) [0x7f66548c8be0]
 5: (ECBackend::handle_message(std::shared_ptr<OpRequest>)+0x186) [0x7f66548d3736]
 6: (ReplicatedPG::do_request(std::shared_ptr<OpRequest>&, ThreadPool::TPHandle&)+0xed) [0x7f6654778e9d]
 7: (OSD::dequeue_op(boost::intrusive_ptr<PG>, std::shared_ptr<OpRequest>, ThreadPool::TPHandle&)+0x3f5) [0x7f665462dbc5]
 8: (PGQueueable::RunVis::operator()(std::shared_ptr<OpRequest> const&)+0x5d) [0x7f665462dded]
 9: (OSD::ShardedOpWQ::_process(unsigned int, ceph::heartbeat_handle_d*)+0x7dd) [0x7f665465449d]
 10: (ShardedThreadPool::shardedthreadpool_worker(unsigned int)+0x947) [0x7f6654c76847]
 11: (ShardedThreadPool::WorkThreadSharded::entry()+0x10) [0x7f6654c789a0]
 12: (()+0x8184) [0x7f665329a184]
 13: (clone()+0x6d) [0x7f66519ee37d]

/home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/test-erasure-eio.sh: line 87:  6804 Terminated              rados --pool $poolname get $objname $dir/COPY
/home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/test-erasure-eio.sh:104: rados_get:  return 1
/home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/test-erasure-eio.sh:227: rados_get_data_bad_size:  return 1
/home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/test-erasure-eio.sh:281: TEST_rados_get_bad_size_shard_0:  return 1
/home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/test-erasure-eio.sh:38: run:  return 1

[tchaikov]

retest this please, jenkins!

[ghost]

jenkins test this please (eio now ignored in master)

[tchaikov]

154/154 Test #143: unittest_pglog ..........................***Timeout 3600.01 sec

[tchaikov]

retest this please

[ghost]

@tchaikov I saw that error a few times. Do you know if there is an issue about it ?

[tchaikov]

@dachary no, seems tracker does not have a ticket tracking it yet.

[ghost]

@tchaikov ack. I'll create one the next time it shows.

[tchaikov]

http://pulpito.ceph.com/kchai-2016-11-22_09:44:32-rados-wip-kefu-testing---basic-mira/